fix: allow include_thoughts + thinking_level together for Gemini 3.x

[hustxiayang]

Description
For GCP Vertex Gemini models, the gateway could not send thinking_level and include_thoughts in the same request. As a result, callers using reasoning_effort (which maps to thinking_level) could never receive thought summaries (reasoning_content), even though Gemini 3.x returns them when both fields are set.

This pr is to allow a separate bool value: include_thoughts to determine whether it returns the thought summary.

client.chat.completions.create(
    model="gemini-3.5-flash",
    messages=[...],
    reasoning_effort="high",                 # depth  -> ThinkingLevel  (unchanged for all models)
    extra_body={"include_thoughts": True},   # opt-in summary -> IncludeThoughts  (new, Gemini-only)
)

This would not break existing interface. There is a bit duplication with the budget setup, but it's fine as the budget setup is deprecating.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 84.85%. Comparing base (e2c59d3) to head (500a7e3).
⚠️ Report is 2 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2433      +/-   ##
==========================================
- Coverage   84.86%   84.85%   -0.02%     
==========================================
  Files         154      154              
  Lines       22419    22423       +4     
==========================================
+ Hits        19026    19027       +1     
- Misses       2237     2239       +2     
- Partials     1156     1157       +1     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.